IETF proposes the Differentiated Service(DiffServ) architecture for next-generation QoS networks. The main features of a DiffServ network are high scalability and compatibility. However, lacking of supporting QoS for individual flows becomes a problem. To remedy this problem, we propose an effective TCP transmission (ETCP) scheme based on the characteristics of TCP. The goal of this thesis is to make TCP transmission more effective. Because of the network complexity, we need to modify both the Ingress and Egress nodes. At Ingress node, we measure TCP congestion window(cwnd) and Round Trip Time(RTT), while at Egress node we monitor the throughput and embed a feedback ratio into TCP header of the acknowledge packets. When the throughput is below or over the target throughput, we adjust DSCP dynamically. The adjustment depends on both measured TCP congestion window and RTT. Each DSCP value is mapped with a delay time through Delayed-FIFO. Changing the delay time in Ingress node, we are able to control TCP transmission associated with RTT more effectively. For the purpose of the demonstration, we implement our scheme on Linux platform. We discuss the performance difference between the theoretical and practical results through experiments. Besides, from the implementation, we have shown that our proposed architecture is workable and compatible with today¡¦s network environment.

In a DiffServ networks, edge and core router classify traffic flows into different PHBs and provide different QoS for the classified flows. In order to achieve satisfactory QoS guarantee, many packet schedulers were proposed. However IETF have not formally standardized an appropriate and effective packet scheduler to minimize the jitter for real-time traffic. In RFC, EF flows are characterized with low-latency, low packet loss rate, and low jitter. Therefore, real-time traffic is often classified into EF flow. By considering the characteristics of real-time traffic, it is not appropriate to forward packets either too fast or too slow. Hence, in this Thesis, we propose a mechanism in which each packet is attached with its own per-hop queuing delay. If a packet is forwarded within its own per-hop queuing delay, we say the packet may arrive too early (credit accumulation). If a packet is forwarded beyond its own per-hop queuing delay, we say the packet has late arrival (deficit accumulation). The Credit/Deficit information can be stored in the IPv6 optional header so that it can pass through the whole networks. If we can minimize the Credit/Deficit, the jitter can be minimized too. Our design is based on a modified WFQ by adding functions such as estimated queuing delay and dynamic class changes. The dynamic class changes allow EF packets to switch among queues to achieve lower jitter and constant delay. We first implement the traditional WFQ scheduler on Linux platform and then followed by the implementation of the Credit/Deficit WFQ (CDWFQ). The experimental results have shown that CDWFQ can provide nearly constant queuing delay, lower packet loss rate, and lower jitter for EF traffic flows.

Video streaming through IP networks has risen rapidly over the recent years, and will continue to do so over the coming years. In addition to this, new technologies such as Virtual Reality and robotics will lead to many new applications that will put high pressure on the networks. To combat these challenges, networks need to be application sensitive, and be able to provide Quality of Service (QoS) based on requirement. Network paradigms like Software Defined Networking (SDN) enables the network to be directly programmable, and could thus solve the challenge. In this thesis, the objective is to research if SDN can provide High QoS. Methods are developed to achieve High QoS with SDN. A combination of Differentiated Services Code Point (DSCP) values and DSCP remarking with Meters are used enable High QoS and together they can give bandwidth guarantee. As a result of the thesis, a solid theoretical method is provided for achieving QoS, tests are performed and show that QoS can be implemented in SDN, but it is unable to implement High QoS due to the lack of implementation for Meters with DSCP remarking.

The doctoral thesis is focused on a research in the area of the quality-of-service support technologies in data networks. The current modern communication networks cannot operate correctly without an effective tool allowing differentiated treatment for various network traffic classes. Looking at the current trends in this area it turns out that the technology of Differentiated services is currently the most widely used mechanism for QoS assurance in data networks. The major part of this doctoral thesis concerns the design of a novel QoS system which offers a solution for one of the main problems of DiffServ technology. This disadvantage lies in the missing cooperation between the end station and edge node of the DiffServ domain. To overcome this limitation the system proposed introduces an improvement which enables the user application to actively participate in the resource reservation process by direct configuration of the DSCP value in the IP header of its own packets. This functionality is based on the identification of DiffServ configuration parameters available in the edge router. To retrieve the information required from network component the well-known SNMP protocol has been chosen, which has direct access to the components’s configuration stored in the MIB database. On the basis of this theoretical proposal several simulation scenarios have been created and analysed. The results show that the system designed presents an efficient solution for the mentioned problem of DiffServ. They also give good assumptions for the successful implementation of this system into a real environment.

This diploma thesis deals with switching components of modern communication networks. A switch architecture focused on the quality of service is described in this paper and there are two switch models designed and simulated in the simulating program MATLAB-SIMULINK. The first model is based on classic switching and the other one is controlled by neural network. A laboratory exercise suitable for communication networks education is created in this paper on the basis of designed models.